Pick the right churn cohort
Do not interview all canceled subscribers as one generic group. Segment by tenure, SKU, order count, cancellation reason, discount exposure, and whether the customer skipped or paused before canceling.
The cohort tells you what kind of decision you can make from the answers. A first-renewal churn cohort, for example, usually surfaces expectation and onboarding issues very differently than a long-tenure churn cohort with stockpile or cadence fatigue.
Run the interview around the timeline
A useful cancellation interview asks what the customer expected when subscribing, what the first order felt like, when the subscription started feeling less useful, and what finally made cancellation feel like the right move.
This timeline reveals earlier intervention points that a cancellation form will usually miss. It also helps teams catch the hidden category errors inside cancel data: what looks like price, taste, or general dissatisfaction may really be overstock, timing friction, or uncertainty about whether the product was working yet.
Choose the retention test
After synthesis, each major theme should become a test. Overstock might become cadence controls. Confusion might become onboarding education. Price concerns might become value proof or bundle framing. Product disappointment might become expectation-setting or product work.
A good playbook ends with decisions, owners, and KPIs. The goal is not just better listening. It is a shorter path from churn evidence to the next shipped retention change.